RTX PRO 6000 Blackwell Max-Q

11 records across 3 models (quants: BF16, NVFP4).

hardwaremodelquantbackendtok/spp tok/sWttft sctx/tgdateclassflagssource
RTX PRO 6000 Blackwell Max-QAgnes-3.0-Flash BF16SGLang125.6––––2026-10-03community–source
RTX PRO 6000 Blackwell Max-QAgnes-3.0-Flash BF16SGLang123.7––––2026-10-03community–source
RTX PRO 6000 Blackwell Max-QAgnes-3.0-Flash BF16SGLang71.6––––2026-10-03community–source
RTX PRO 6000 Blackwell Max-QAgnes-3.0-Flash BF16SGLang59.0––––2026-10-03community–source
RTX PRO 6000 Blackwell Max-QAgnes-3.0-Flash BF16SGLang35.2––––2026-10-03community–source
RTX PRO 6000 Blackwell Max-QQwen3.5-35B-A3B 35B MoE A3BBF16vLLM3341.89––––2026-02community–source
RTX PRO 6000 Blackwell Max-QQwen3.8-27B-TURBO NVFP4vLLM765.6––––2026-10-03community–source
RTX PRO 6000 Blackwell Max-QQwen3.8-27B-TURBO NVFP4vLLM581.0––––2026-10-03community–source
RTX PRO 6000 Blackwell Max-QQwen3.8-27B-TURBO NVFP4vLLM274.5––––2026-10-03community–source
RTX PRO 6000 Blackwell Max-QQwen3.8-27B-TURBO NVFP4vLLM209.8––––2026-10-03community–source
RTX PRO 6000 Blackwell Max-QQwen3.8-27B-TURBO NVFP4vLLM106.3––––2026-10-03community–source

All rows cite their source. Provenance: community 40018 sourced 1085

← all hardware